Linear Discriminant Analysis (LDA) More...

#include <shark/Algorithms/Trainers/LDA.h>

Inheritance diagram for shark::LDA:

Public Member Functions
	LDA (double regularization=0.0)
	constructor

std::string	name () const
	From INameable: return the class name.

double	regularization () const
	return the regularization constant

void	setRegularization (double regularization)
	set the regularization constant. 0 means no regularization.

RealVector	parameterVector () const
	inherited from IParameterizable; read the regularization parameter

void	setParameterVector (RealVector const &param)
	inherited from IParameterizable; set the regularization parameter

size_t	numberOfParameters () const
	inherited from IParameterizable

SHARK_EXPORT_SYMBOL void	train (LinearClassifier<> &model, LabeledData< RealVector, unsigned int > const &dataset)
	Compute the LDA solution for a multi-class problem.

SHARK_EXPORT_SYMBOL void	train (LinearClassifier<> &model, WeightedLabeledData< RealVector, unsigned int > const &dataset)
	Compute the LDA solution for a weighted multi-class problem.

Public Member Functions inherited from shark::AbstractWeightedTrainer< LinearClassifier<>, unsigned int >
virtual void	train (ModelType &model, WeightedDatasetType const &dataset)=0
	Executes the algorithm and trains a model on the given weighted data.

virtual void	train (ModelType &model, DatasetType const &dataset)
	Executes the algorithm and trains a model on the given unweighted data.

Public Member Functions inherited from shark::AbstractTrainer< Model, LabelTypeT >
virtual void	train (ModelType &model, DatasetType const &dataset)=0
	Core of the Trainer interface.

Public Member Functions inherited from shark::INameable
virtual	~INameable ()

Public Member Functions inherited from shark::ISerializable
virtual	~ISerializable ()
	Virtual d'tor.

virtual void	read (InArchive &archive)
	Read the component from the supplied archive.

virtual void	write (OutArchive &archive) const
	Write the component to the supplied archive.

void	load (InArchive &archive, unsigned int version)
	Versioned loading of components, calls read(...).

void	save (OutArchive &archive, unsigned int version) const
	Versioned storing of components, calls write(...).

	BOOST_SERIALIZATION_SPLIT_MEMBER ()

Public Member Functions inherited from shark::IParameterizable< VectorType >
virtual	~IParameterizable ()

Protected Attributes
double	m_regularization

Additional Inherited Members
Public Types inherited from shark::AbstractWeightedTrainer< LinearClassifier<>, unsigned int >
typedef base_type::ModelType	ModelType

typedef base_type::InputType	InputType

typedef base_type::LabelType	LabelType

typedef base_type::DatasetType	DatasetType

typedef WeightedLabeledData< InputType, LabelType >	WeightedDatasetType

Public Types inherited from shark::AbstractTrainer< Model, LabelTypeT >
typedef Model	ModelType

typedef ModelType::InputType	InputType

typedef LabelTypeT	LabelType

typedef LabeledData< InputType, LabelType >	DatasetType

Public Types inherited from shark::IParameterizable< VectorType >
typedef VectorType	ParameterVectorType

Detailed Description

Linear Discriminant Analysis (LDA)

This classes implements the well known linear discriminant analysis. LDA assumes that every point is drawn from a multivariate normal distributions. Every class has its own mean but all classes have the same covariance.

An arbitrary number of classes is supported. The resulting model is of the form

\[ \arg \max_c \log(p(x|c)*P(c)) \]

where \( p(x|c) = \exp(-(x-m_c)^T(C+\alpha I)(x-m_c)) \). \( m_c\) are the means of class c, \( C \) is the covariance matrix formed by all data points. The regularization paramter \( \alpha \) is by default 0. The trainer is implemented such, that it still works when C is singular, in this case the singular directions are ignored.

Definition at line 59 of file LDA.h.

Constructor & Destructor Documentation

◆ LDA()

shark::LDA::LDA ( double regularization = 0.0 )

inline

constructor

Definition at line 63 of file LDA.h.

References regularization(), and setRegularization().

Member Function Documentation

◆ name()

std::string shark::LDA::name ( ) const

inlinevirtual

From INameable: return the class name.

Reimplemented from shark::INameable.

Definition at line 68 of file LDA.h.

◆ numberOfParameters()

size_t shark::LDA::numberOfParameters ( ) const

inlinevirtual

inherited from IParameterizable

Reimplemented from shark::IParameterizable< VectorType >.

Definition at line 94 of file LDA.h.

◆ parameterVector()

RealVector shark::LDA::parameterVector ( ) const

inlinevirtual

inherited from IParameterizable; read the regularization parameter

Reimplemented from shark::IParameterizable< VectorType >.

Definition at line 83 of file LDA.h.

References m_regularization.

◆ regularization()

double shark::LDA::regularization ( ) const

inline

return the regularization constant

Definition at line 72 of file LDA.h.

References m_regularization.

Referenced by LDA(), and setRegularization().

◆ setParameterVector()

void shark::LDA::setParameterVector ( RealVector const & param )

inlinevirtual

inherited from IParameterizable; set the regularization parameter

Reimplemented from shark::IParameterizable< VectorType >.

Definition at line 89 of file LDA.h.

References m_regularization, and SIZE_CHECK.

◆ setRegularization()

void shark::LDA::setRegularization ( double regularization )

inline

set the regularization constant. 0 means no regularization.

Definition at line 77 of file LDA.h.

References m_regularization, RANGE_CHECK, and regularization().

Referenced by LDA().

◆ train() [1/2]

SHARK_EXPORT_SYMBOL void shark::LDA::train	(	LinearClassifier<> &	model,
		LabeledData< RealVector, unsigned int > const &	dataset
	)

Compute the LDA solution for a multi-class problem.

Referenced by main().

◆ train() [2/2]

SHARK_EXPORT_SYMBOL void shark::LDA::train	(	LinearClassifier<> &	model,
		WeightedLabeledData< RealVector, unsigned int > const &	dataset
	)

Compute the LDA solution for a weighted multi-class problem.

Member Data Documentation

◆ m_regularization

double shark::LDA::m_regularization

protected

The regularization parameter \( \lambda \) adds \( - \lambda I \) to the second moment matrix, where \( I \) is the identity matrix

Definition at line 107 of file LDA.h.

Referenced by parameterVector(), regularization(), setParameterVector(), and setRegularization().

The documentation for this class was generated from the following file:

include/shark/Algorithms/Trainers/LDA.h

Public Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ LDA()

Member Function Documentation

◆ name()

◆ numberOfParameters()

◆ parameterVector()

◆ regularization()

◆ setParameterVector()

◆ setRegularization()

◆ train() [1/2]

◆ train() [2/2]

Member Data Documentation

◆ m_regularization